ETTR (Expose To The Right)

[koraks]

My interpretation of the statement, as made..."The lightmeter assumes you read Zone V (middle tone) brightness area of the subject; so if you meter the 'bright area with detail' of the scene (which should fall into Zone VII) you need to adjust the indicated exposure by -2EV to expose so that detail is retained." Clear to me, not confusing, about the principles of exposure (although not the accompanying darkroom usage) of Zone system.

Well, if that's what it is, then the approach will fail as it'll result in horribly clipped highlights. If you place the highlights that are supposed to fall on VII in the final image at +2EV exposure, most of zones VIII and IX will be clipped. I'm sure that's not what the author of the book intends to convey, but it's precisely why I mentioned that this unnecessarily (and perhaps dangerously) confuses things. I'm sure that reading the entire book would resolve the matter, but then again, how does that fit with OP's desire to keep thing simple? I find it risky to try and marry a digital-interpreted zone system approach with ETTR when someone is looking for a simple approach.

 

[wiltw]

Well, if that's what it is, then the approach will fail as it'll result in horribly clipped highlights. If you place the highlights that are supposed to fall on VII in the final image at +2EV exposure, most of zones VIII and IX will be clipped. I'm sure that's not what the author of the book intends to convey, but it's precisely why I mentioned that this unnecessarily (and perhaps dangerously) confuses things. I'm sure that reading the entire book would resolve the matter, but then again, how does that fit with OP's desire to keep thing simple? I find it risky to try and marry a digital-interpreted zone system approach with ETTR when someone is looking for a simple approach.

Not arguing, but explaining...

1. Things in Zone VIII of the scene are by definition 'with texture' (or 'with detail'), while Zone IX is 'without texture, approaching pure white'
2. If you meter something with Zone VII brightness, it is +2EV brighter than Zone V (middletone), so to record it properly you adjust what the meter suggests (while pointed at Zone VII), by -2EV
   things inherently in Zone VII would never be 'without texture, approaching pure white' unless it was recorded with exposure +2EV-- bringing those areas into Zone IX

The concept is simple..."If you meter something brighter than middle tone, to you need to compensate the indicated meter reading in the opposite direction to record it at its inherent brightness"
that concept is fundamental to why you adjust meter readings of snow scenes by using exposure compensation to bring the indicated exposure to the actual exposure (so as not to record 'gray snow'

 

[koraks]

Not arguing, but explaining...

That's the added value of all this!

Things in Zone VIII of the scene are by definition 'with texture' (or 'with detail'), while Zone IX is 'without texture, approaching pure white'

To have tone, on digital you need to steer clear of channel clipping. Since clipping generally happens on one channel first, inching close to that point is really risky. To get tone for zone IX, no clipping may occur on any channel as you generally run into really severe color aberrations otherwise.

I'm sure you've played with a digital camera enough to know how far you can push the overexposure of important texture. Depending on the camera make & model, subject matter, metering pattern etc., things break down starting around +2EV. This is really an area where digital behaves fundamentally differently from film; the latter has graceful degradation, the former doesn't award us that luxury.

 

[wiltw]

That's the added value of all this!


To have tone, on digital you need to steer clear of channel clipping. Since clipping generally happens on one channel first, inching close to that point is really risky. To get tone for zone IX, no clipping may occur on any channel as you generally run into really severe color aberrations otherwise.

I'm sure you've played with a digital camera enough to know how far you can push the overexposure of important texture. Depending on the camera make & model, subject matter, metering pattern etc., things break down starting around +2EV. This is really an area where digital behaves fundamentally differently from film; the latter has graceful degradation, the former doesn't award us that luxury.

Here is such an illustration, shot many years ago...
I put a piece of while coffee filter paper in the sun, sprinkled something white (salt?/sugar?) onto it, and then recorded a series of exposures. The first shot renders the middle tone (18% gray exposure card) at its inherent brightness, or 'proper exposure'. Then I recorded each successive shot +1EV brighter than that prior shot, concluding with a shot -1EV from 'proper' exposure.



...notice that the white substance disappeared at +3EV, and was scarcely detectable at +2EV although the filter paper was still easily discernable from the 18% card, and (-- except for the shadows cast -- ) it had disappeared into the card at +3EV

 

[koraks]

Frame #1 probably already lacks texture in the white particulate matter, #2 is exposed into oblivion. This illustrates my point - while your 18% grey card may come out OK at +2EV, the problem is that the world we photograph is generally not a featureless field of even tone. It has texture, and texture means brightness variation. Things start to break down pretty quickly if you err to the side of overexposure. Even if you spot meter with a narrow 1% angle (or so, whatever your digicam affords), many textures are finer than that measurement angle - and they're by far not always specular highlights you'd want to lose.

What I notice is that essays that argue for ETTR often rely on very flat scenes with very little brightness variation. I agree that in such cases, you could overexpose a bit to optimize s/n ratio. The real-world gain in image quality will be really, really limited. Overall - not worth it, IMO.

What is worth it in my book is to determine how much overexposure your digital camera can take before things start to go south and then ensure you steer clear of that point for anything that needs tone or texture. Which is very much like how you'd shoot slide film.

 

[Alan Edward Klein]

The idea is that you get marginally better signal to noise ratio by shifting the entire histogram as much to the right as possible. But the risk is of course highlight clipping, which makes this a tool that needs to be applied with care. I personally find it very doubtful that the net benefit is worth it in practice.

Especially today, with newer cameras having more dynamic range.

 

[wiltw]

Frame #1 probably already lacks texture in the white particulate matter, #2 is exposed into oblivion. This illustrates my point - while your 18% grey card may come out OK at +2EV, the problem is that the world we photograph is generally not a featureless field of even tone. It has texture, and texture means brightness variation. Things start to break down pretty quickly if you err to the side of overexposure. Even if you spot meter with a narrow 1% angle (or so, whatever your digicam affords), many textures are finer than that measurement angle - and they're by far not always specular highlights you'd want to lose.

What I notice is that essays that argue for ETTR often rely on very flat scenes with very little brightness variation. I agree that in such cases, you could overexpose a bit to optimize s/n ratio. The real-world gain in image quality will be really, really limited. Overall - not worth it, IMO.

What is worth it in my book is to determine how much overexposure your digital camera can take before things start to go south and then ensure you steer clear of that point for anything that needs tone or texture. Which is very much like how you'd shoot slide film.

Again, no argument. This is getting 'into the weeds'.
I will point out that if one is USING the histogram on the camera to keep pixels from falling off the edge of the histogram, the inherent definition of ETTR, then detail is preserved! So then rather than simply -2EV when reading something with Zone VII brightness, one might be actually using -2.33EV and preserving detail. The 'read Zone VI, then adjust -2EV' was, I believe, for discussion of the fundamental principle.

BTW, I just looked at the original shots and in the 'proper exposure' (shot 1) the white substance was already close to the right edge of the histogram (falling into Zone VIII brightness)
\

so at +2EV it was indeed 'off the histogram' and 'exposed into oblivion'!

 

[Nikon 2]

Again, no argument. This is getting 'into the weeds'.
I will point out that if one is USING the histogram on the camera to keep pixels from falling off the edge of the histogram, the inherent definition of ETTR, then detail is preserved! So then rather than simply -2EV when reading something with Zone VII brightness, one might be actually using -2.33EV and preserving detail. The 'read Zone VI, then adjust -2EV' was, I believe, for discussion of the fundamental principle.

I’m glad it’s not me…

 

[wiltw]

I’m glad it’s not me…

With regard to the initial 'muddy' shots of yours, I believe that your metering was wrongly biased by the bright sky (and the big white clouds). Folks often assume it is 0EV, when in fact it starts off brighter than an 18% gray card in the same light. Today, my sky (without clouds) is not exceptionally bright, but it does read +0.4EV brighter than a gray card right now.

 

[koraks]

if one is USING the histogram on the camera to keep pixels from falling off the edge of the histogram, the inherent definition of ETTR, then detail is preserved!

Yes, absolutely; that's of course also the advantage of digital. And agree that we are/were getting in the thick of it. I went down that road 15 or so years ago, took the theory into the field and then decided on the basis of that experience the benefit didn't outweigh the risk. Today when I shoot digital I meter in various ways, but always ensure that I'm not lopping off any important highlight tonality. Then figure out the rest in post!

 

[Nikon 2]

With regard to the initial 'muddy' shots of yours, I believe that your metering was wrongly biased by the bright sky (and the big white clouds). Folks often assume it is 0EV, when in fact it starts off brighter than an 18% gray card in the same light. Today, my sky (without clouds) is not exceptionally bright, but it does read +0.4EV brighter than a gray card right now.

How do you read metering on your camera…